Pour voir les autres types de publications sur ce sujet consultez le lien suivant : AI Generated Text Detection.
Articles de revues sur le sujet « AI Generated Text Detection »
Créez une référence correcte selon les styles APA, MLA, Chicago, Harvard et plusieurs autres
Consultez les 50 meilleurs articles de revues pour votre recherche sur le sujet « AI Generated Text Detection ».
À côté de chaque source dans la liste de références il y a un bouton « Ajouter à la bibliographie ». Cliquez sur ce bouton, et nous générerons automatiquement la référence bibliographique pour la source choisie selon votre style de citation préféré : APA, MLA, Harvard, Vancouver, Chicago, etc.
Vous pouvez aussi télécharger le texte intégral de la publication scolaire au format pdf et consulter son résumé en ligne lorsque ces informations sont inclues dans les métadonnées.
Parcourez les articles de revues sur diverses disciplines et organisez correctement votre bibliographie.
1
Bhattacharjee, Amrita, et Huan Liu. « Fighting Fire with Fire : Can ChatGPT Detect AI-generated Text ? » ACM SIGKDD Explorations Newsletter 25, no 2 (26 mars 2024) : 14–21. http://dx.doi.org/10.1145/3655103.3655106.
Texte intégralRésumé :
Large language models (LLMs) such as ChatGPT are increasingly being used for various use cases, including text content generation at scale. Although detection methods for such AI-generated text exist already, we investigate ChatGPT's performance as a detector on such AI-generated text, inspired by works that use ChatGPT as a data labeler or annotator. We evaluate the zeroshot performance of ChatGPT in the task of human-written vs. AI-generated text detection, and perform experiments on publicly available datasets. We empirically investigate if ChatGPT is symmetrically effective in detecting AI-generated or human-written text. Our findings provide insight on how ChatGPT and similar LLMs may be leveraged in automated detection pipelines by simply focusing on solving a specific aspect of the problem and deriving the rest from that solution. All code and data is available at https://github.com/AmritaBh/ChatGPT-as-Detector.
2
Wang, Yu. « Survey for Detecting AI-generated Content ». Advances in Engineering Technology Research 11, no 1 (18 juillet 2024) : 643. http://dx.doi.org/10.56028/aetr.11.1.643.2024.
Texte intégralRésumé :
In large language models (LLMs) field, the rapid advancements have significantly improved text generation, which has blured the distinction between AI-generated and human-written texts. These developments have sparked concerns about potential risks, such as disseminating fake information or engaging in academic cheating. As the responsible use of LLMs becomes imperative, the detection of AI-generated content has become a crucial task. Most existing surveys on AI-generated text (AIGT) Detection have analysed the detection approaches from a computational perspective, with less attention to linguistic aspects. This survey seeks to provide a fresh perspective to drive progress in the area of LLM-generated text detection. Futhermore, in order to make the assessment more explainable, we emphasize the great importence of leveraging specific parameters or metrics to linguistically evaluate the candidate text.
3
A, Nykonenko. « How Text Transformations Affect AI Detection ». Artificial Intelligence 29, AI.2024.29(4) (30 décembre 2024) : 233–41. https://doi.org/10.15407/jai2024.04.233.
Texte intégralRésumé :
This study addresses the critical issue of AI writing detection, which currently plays a key role in deterring technology misuse and proposes a foundation for the controllable and conscious use of AI. The ability to differentiate between human-written and AI-generated text is crucial for the practical application of any policies or guidelines. Current detection tools are unable to interpret their decisions in a way that is understandable to humans or provide any human-readable evidence or proof for their decisions. We assume that there should be a traceable footprint in LLM-generated texts that is invisible to the human eye but can be detected by AI detection tools-referred to as the AI footprint. Understanding its nature will help bring more light into the guiding principles lying at the core of AI detection technology and help build more trust in the technology in general. The main goal of this paper is to examine the AI footprint in text data generated by large language models (LLMs). To achieve this, we propose a new method for text transformation that should measurably decrease the AI footprint in the text data, impacting AI writing scores. We applied a set of stage-by-stage text transformations focused on decreasing meaningfulness by masking or removing words. Using a set of AI detectors, we measured the AI writing score as a proxy metric for assessing the impact of the proposed method. The results demonstrate a significant correlation between the severity of changes and the resulting impact on AI writing scores, highlighting the need for developing more reliable AI writing identification methods that are immune to attempts to hide the AI footprint through subtle changes
4
Singh, Dr Viomesh, Bhavesh Agone, Aryan More, Aryan Mengawade, Atharva Deshmukh et Atharva Badgujar. « SAVANA- A Robust Framework for Deepfake Video Detection and Hybrid Double Paraphrasing with Probabilistic Analysis Approach for AI Text Detection ». International Journal for Research in Applied Science and Engineering Technology 12, no 11 (30 novembre 2024) : 2074–83. http://dx.doi.org/10.22214/ijraset.2024.65526.
Texte intégralRésumé :
Abstract: As the generative AI has advanced with a great speed, the need to detect AI-generated content, including text and deepfake media, also increased. This research work proposes a hybrid detection method that includes double paraphrasing-based consistency checks, coupled with probabilistic content analysis through natural language processing and machine learning algorithms for text and advanced deepfake detection techniques for media. Our system hybridizes the double paraphrasing framework of SAVANA with probabilistic analysis toward high accuracy on AI-text detection in forms such as DOCX or PDF from diverse domains- academic text, business text, reviews, and media. Specifically, for detecting visual artifact and spatiotemporal inconsistencies attributed to deepfakes within media applications, we'll be exploiting BlazeFace, EfficientNetB4 for extracting features while classifying and detecting respective deepfakes. Experimental results indicate that the hybrid model achieves up to 95% accuracy for AI-generated text detection and up to 96% accuracy for deepfake detection with the traditional models and the standalone SAVANA-based methods. This approach therefore positions our framework as an adaptive and reliable tool to detect AI-generated content within various contexts, thereby enriching content integrity in digital environments.
5
Vismay Vora, Et al. « A Multimodal Approach for Detecting AI Generated Content using BERT and CNN ». International Journal on Recent and Innovation Trends in Computing and Communication 11, no 9 (30 octobre 2023) : 691–701. http://dx.doi.org/10.17762/ijritcc.v11i9.8861.
Texte intégralRésumé :
With the advent of Generative AI technologies like LLMs and image generators, there will be an unprecedented rise in synthetic information which requires detection. While deepfake content can be identified by considering biological cues, this article proposes a technique for the detection of AI generated text using vocabulary, syntactic, semantic and stylistic features of the input data and detecting AI generated images through the use of a CNN model. The performance of these models is also evaluated and benchmarked with other comparative models. The ML Olympiad Competition dataset from Kaggle is used in a BERT Model for text detection and the CNN model is trained on the CIFAKE dataset to detect AI generated images. It can be concluded that in the upcoming era, AI generated content will be omnipresent and no single model will truly be able to detect all AI generated content especially when these technologies are getting better.
6
Subramaniam, Raghav. « Identifying Text Classification Failures in Multilingual AI-Generated Content ». International Journal of Artificial Intelligence & ; Applications 14, no 5 (28 septembre 2023) : 57–63. http://dx.doi.org/10.5121/ijaia.2023.14505.
Texte intégralRésumé :
With the rising popularity of generative AI tools, the nature of apparent classification failures by AI content detection softwares, especially between different languages, must be further observed. This paper aims to do this through testing OpenAI’s “AI Text Classifier” on a set of human and AI-generated texts inEnglish, German, Arabic, Hindi, Chinese, and Swahili. Given the unreliability of existing tools for detection of AIgenerated text, it is notable that specific types of classification failures often persist in slightly different ways when various languages are observed: misclassification of human-written content as “AI-generated” and vice versa may occur more frequently in specific language content than others. Our findings indicate that false negative labelings are more likely to occur in English, whereas false positives are more likely to occur in Hindi and Arabic. There was an observed tendency for other languages to not be confidently labeled at all.
7
Sushma D S, Pooja C N, Varsha H S, Yasir Hussain et P Yashash. « Detection and Classification of ChatGPT Generated Contents Using Deep Transformer Models ». International Research Journal on Advanced Engineering Hub (IRJAEH) 2, no 05 (23 mai 2024) : 1404–7. http://dx.doi.org/10.47392/irjaeh.2024.0193.
Texte intégralRésumé :
AI advancements, particularly in neural networks, have brought about groundbreaking tools like text generators and chatbots. While these technologies offer tremendous benefits, they also pose serious risks such as privacy breaches, spread of misinformation, and challenges to academic integrity. Previous efforts to distinguish between human and AI-generated text have been limited, especially with models like ChatGPT. To tackle this, we created a dataset containing both human and ChatGPT-generated text, using it to train and test various machine and deep learning models. Your results, particularly the high F1-score and accuracy achieved by the RoBERTa-based custom deep learning model and Distil BERT, indicate promising progress in this area. By establishing a robust baseline for detecting and classifying AI-generated content, your work contributes significantly to mitigating potential misuse of AI-powered text generation tools.
8
Alshammari, Hamed, et Khaled Elleithy. « Toward Robust Arabic AI-Generated Text Detection : Tackling Diacritics Challenges ». Information 15, no 7 (19 juillet 2024) : 419. http://dx.doi.org/10.3390/info15070419.
Texte intégralRésumé :
Current AI detection systems often struggle to distinguish between Arabic human-written text (HWT) and AI-generated text (AIGT) due to the small marks present above and below the Arabic text called diacritics. This study introduces robust Arabic text detection models using Transformer-based pre-trained models, specifically AraELECTRA, AraBERT, XLM-R, and mBERT. Our primary goal is to detect AIGTs in essays and overcome the challenges posed by the diacritics that usually appear in Arabic religious texts. We created several novel datasets with diacritized and non-diacritized texts comprising up to 9666 HWT and AIGT training examples. We aimed to assess the robustness and effectiveness of the detection models on out-of-domain (OOD) datasets to assess their generalizability. Our detection models trained on diacritized examples achieved up to 98.4% accuracy compared to GPTZero’s 62.7% on the AIRABIC benchmark dataset. Our experiments reveal that, while including diacritics in training enhances the recognition of the diacritized HWTs, duplicating examples with and without diacritics is inefficient despite the high accuracy achieved. Applying a dediacritization filter during evaluation significantly improved model performance, achieving optimal performance compared to both GPTZero and the detection models trained on diacritized examples but evaluated without dediacritization. Although our focus was on Arabic due to its writing challenges, our detector architecture is adaptable to any language.
9
Jeremie Busio Legaspi, Roan Joyce Ohoy Licuben, Emmanuel Alegado Legaspi et Joven Aguinaldo Tolentino. « Comparing ai detectors : evaluating performance and efficiency ». International Journal of Science and Research Archive 12, no 2 (30 juillet 2024) : 833–38. http://dx.doi.org/10.30574/ijsra.2024.12.2.1276.
Texte intégralRésumé :
The widespread utilization of AI tools such as ChatGPT has become increasingly prevalent among learners, posing a threat to academic integrity. This study seeks to evaluate capability and efficiency of AI detection tools in distinguishing between human-authored and AI-generated works. Three-paragraph works on “AutoCAD and Architecture” were generated through ChatGPT, and three human-written works were subjected to evaluation. AI detection tools such as GPTZero, Copyleaks and Writer AI were used to evaluate these paragraphs. Parameters such as “Human/Human Text/Human Generated Text” and “AI/AI Content Detected” were used to evaluate the performance of the three AI detection tools in evaluating outputs. Findings indicate that GPT Zero and Copyleaks have higher reliability in determining human-authored work and AI generated work while Writer AI showed a notable content classification of “Human Generated Content” on all tested outputs showing less sensitivity on determining human-authored work and AI generated work. Findings indicate that the use of Artificial Intelligence as an AI detection tool should be accompanied with thorough validation and cross-referencing of results.
10
Kim, Min-Gyu, et Heather Desaire. « Detecting the Use of ChatGPT in University Newspapers by Analyzing Stylistic Differences with Machine Learning ». Information 15, no 6 (25 mai 2024) : 307. http://dx.doi.org/10.3390/info15060307.
Texte intégralRésumé :
Large language models (LLMs) have the ability to generate text by stringing together words from their extensive training data. The leading AI text generation tool built on LLMs, ChatGPT, has quickly grown a vast user base since its release, but the domains in which it is being heavily leveraged are not yet known to the public. To understand how generative AI is reshaping print media and the extent to which it is being implemented already, methods to distinguish human-generated text from that generated by AI are required. Since college students have been early adopters of ChatGPT, we sought to study the presence of generative AI in newspaper articles written by collegiate journalists. To achieve this objective, an accurate AI detection model is needed. Herein, we analyzed university newspaper articles from different universities to determine whether ChatGPT was used to write or edit the news articles. We developed a detection model using classical machine learning and used the model to detect AI usage in the news articles. The detection model showcased a 93% accuracy in the training data and had a similar performance in the test set, demonstrating effectiveness in AI detection above existing state-of-the-art detection tools. Finally, the model was applied to the task of searching for generative AI usage in 2023, and we found that ChatGPT was not used to revise articles to any appreciable measure to write university news articles at the schools we studied.
11
Wang, Hao, Jianwei Li et Zhengyu Li. « AI-generated text detection and classification based on BERT deep learning algorithm ». Theoretical and Natural Science 39, no 1 (31 juillet 2024) : None. http://dx.doi.org/10.54254/2753-8818/39/20240625.
Texte intégralRésumé :
With the rapid development and wide application of deep learning technology, AI-generated text detection plays an increasingly important role in various fields. In this study, we developed an efficient AI-generated text detection model based on the BERT algorithm, which provides new ideas and methods for solving related problems. In the data preprocessing stage, a series of steps were taken to process the text, including operations such as converting to lowercase, word splitting, removing stop words, stemming extraction, removing digits, and eliminating redundant spaces, to ensure data quality and accuracy. By dividing the dataset into a training set and a test set in the ratio of 60% and 40%, and observing the changes in the accuracy and loss values during the training process, we found that the model performed well during the training process. The accuracy increases steadily from the initial 94.78% to 99.72%, while the loss value decreases from 0.261 to 0.021 and converges gradually, which indicates that the BERT model is able to detect AI-generated text with high accuracy and the prediction results are gradually approaching the real classification results. Further analysis of the results of the training and test sets reveals that in terms of loss value, the average loss of the training set is 0.0565, while the average loss of the test set is 0.0917, showing a slightly higher loss value. As for the accuracy, the average accuracy of the training set reaches 98.1%, while the average accuracy of the test set is 97.71%, which is not much different from each other, indicating that the model has good generalisation ability. In conclusion, the AI-generated text detection model based on the BERT algorithm proposed in this study shows high accuracy and stability in experiments, providing an effective solution for related fields. In the future, the model performance can be further optimised and its potential for application in a wider range of fields can be explored to promote the development and application of AI technology in the field of text detection.
12
Corizzo, Roberto, et Sebastian Leal-Arenas. « One-Class Learning for AI-Generated Essay Detection ». Applied Sciences 13, no 13 (5 juillet 2023) : 7901. http://dx.doi.org/10.3390/app13137901.
Texte intégralRésumé :
Detection of AI-generated content is a crucially important task considering the increasing attention towards AI tools, such as ChatGPT, and the raised concerns with regard to academic integrity. Existing text classification approaches, including neural-network-based and feature-based methods, are mostly tailored for English data, and they are typically limited to a supervised learning setting. Although one-class learning methods are more suitable for classification tasks, their effectiveness in essay detection is still unknown. In this paper, this gap is explored by adopting linguistic features and one-class learning models for AI-generated essay detection. Detection performance of different models is assessed in different settings, where positively labeled data, i.e., AI-generated essays, are unavailable for model training. Results with two datasets containing essays in L2 English and L2 Spanish show that it is feasible to accurately detect AI-generated essays. The analysis reveals which models and which sets of linguistic features are more powerful than others in the detection task.
13
Zeng, Zijie, Lele Sha, Yuheng Li, Kaixun Yang, Dragan Gašević et Guangliang Chen. « Towards Automatic Boundary Detection for Human-AI Collaborative Hybrid Essay in Education ». Proceedings of the AAAI Conference on Artificial Intelligence 38, no 20 (24 mars 2024) : 22502–10. http://dx.doi.org/10.1609/aaai.v38i20.30258.
Texte intégralRésumé :
The recent large language models (LLMs), e.g., ChatGPT, have been able to generate human-like and fluent responses when provided with specific instructions. While admitting the convenience brought by technological advancement, educators also have concerns that students might leverage LLMs to complete their writing assignments and pass them off as their original work. Although many AI content detection studies have been conducted as a result of such concerns, most of these prior studies modeled AI content detection as a classification problem, assuming that a text is either entirely human-written or entirely AI-generated. In this study, we investigated AI content detection in a rarely explored yet realistic setting where the text to be detected is collaboratively written by human and generative LLMs (termed as hybrid text for simplicity). We first formalized the detection task as identifying the transition points between human-written content and AI-generated content from a given hybrid text (boundary detection). We constructed a hybrid essay dataset by partially and randomly removing sentences from the original student-written essays and then instructing ChatGPT to fill in for the incomplete essays. Then we proposed a two-step detection approach where we (1) separated AI-generated content from human-written content during the encoder training process; and (2) calculated the distances between every two adjacent prototypes (a prototype is the mean of a set of consecutive sentences from the hybrid text in the embedding space) and assumed that the boundaries exist between the two adjacent prototypes that have the furthest distance from each other. Through extensive experiments, we observed the following main findings: (1) the proposed approach consistently outperformed the baseline methods across different experiment settings; (2) the encoder training process (i.e., step 1 of the above two-step approach) can significantly boost the performance of the proposed approach; (3) when detecting boundaries for single-boundary hybrid essays, the proposed approach could be enhanced by adopting a relatively large prototype size (i.e., the number of sentences needed to calculate a prototype), leading to a 22% improvement (against the best baseline method) in the In-Domain evaluation and an 18% improvement in the Out-of-Domain evaluation.
14
Krawczyk, Natalia, Barbara Probierz et Jan Kozak. « Towards AI-Generated Essay Classification Using Numerical Text Representation ». Applied Sciences 14, no 21 (26 octobre 2024) : 9795. http://dx.doi.org/10.3390/app14219795.
Texte intégralRésumé :
The detection of essays written by AI compared to those authored by students is increasingly becoming a significant issue in educational settings. This research examines various numerical text representation techniques to improve the classification of these essays. Utilizing a diverse dataset, we undertook several preprocessing steps, including data cleaning, tokenization, and lemmatization. Our system analyzes different text representation methods such as Bag of Words, TF-IDF, and fastText embeddings in conjunction with multiple classifiers. Our experiments showed that TF-IDF weights paired with logistic regression reached the highest accuracy of 99.82%. Methods like Bag of Words, TF-IDF, and fastText embeddings achieved accuracies exceeding 96.50% across all tested classifiers. Sentence embeddings, including MiniLM and distilBERT, yielded accuracies from 93.78% to 96.63%, indicating room for further refinement. Conversely, pre-trained fastText embeddings showed reduced performance, with a lowest accuracy of 89.88% in logistic regression. Remarkably, the XGBoost classifier delivered the highest minimum accuracy of 96.24%. Specificity and precision were above 99% for most methods, showcasing high capability in differentiating between student-created and AI-generated texts. This study underscores the vital role of choosing dataset-specific text representations to boost classification accuracy.
15
Howard, Frederick Matthew, Anran Li, Mark Riffon, Elizabeth Garrett-Mayer et Alexander T. Pearson. « Artificial intelligence (AI) content detection in ASCO scientific abstracts from 2021 to 2023. » Journal of Clinical Oncology 42, no 16_suppl (1 juin 2024) : 1565. http://dx.doi.org/10.1200/jco.2024.42.16_suppl.1565.
Texte intégralRésumé :
1565 Background: Generative AI models such as OpenAI’s ChatGPT have been broadly utilized throughout the medical literature. Previous studies have found that AI can generate scientific abstracts which can be difficult to distinguish from the work of human authors. There is a pressing need to characterize utilization of AI in scientific writing to guide policy. Methods: In collaboration with ASCO's Center for Research and Analytics, we extracted text from all scientific abstracts submitted to ASCO 2021 – 2023 Annual Meetings. Likelihood of AI content was evaluated by four AI detectors: GPTZero, Originality.ai (OGAI), Sapling, and Kashyap's AI Content Detector (AICD). Each detector produces a numeric likelihood of AI content. Predictions were dichotomized for uniform analysis across detectors: those in the top 5% for a given detector were considered as having high likelihood of AI content. Logistic regression was used to compute odds ratio (OR) for AI-generated content with respect to submission year and abstract characteristics. Predictions were also assessed for 10 human-written abstracts as negative controls, and 10 produced by OpenAI’s GPT-3 and GPT-4 models as positive controls. Results: 15,553 abstracts met inclusion criteria. 5,179 (33%) were published online only, and 5,327 (34%) referenced registered clinical trials. Across all detectors, abstracts submitted in 2023 were significantly more likely to contain AI content than those in 2021 (OR range 1.3 - 1.7). In abstracts from 2023, AI content score was associated with online only publication, lack of clinical trial number, and abstract track (Table). None of the 10 negative control human written abstracts were identified as AI generated, whereas 100%, 95%, 90%, and 30% of the GPT-3/4 generated abstracts were classified as AI generated by Sapling, GPTZero, OGAI, and AICD respectively using the 5% threshold. Additional results will be presented. Conclusions: AI content detectors uniformly suggest a higher likelihood of AI content generation for abstracts submitted in 2023. Predicted AI content is associated with triage of abstracts to online only presentation, suggesting predicted AI content is associated with lower perceived abstract quality. Further work is needed to understand the accuracy of AI detectors and utility in the abstract review process. [Table: see text]
16
Xu, Zhenyu, et Victor S. Sheng. « Detecting AI-Generated Code Assignments Using Perplexity of Large Language Models ». Proceedings of the AAAI Conference on Artificial Intelligence 38, no 21 (24 mars 2024) : 23155–62. http://dx.doi.org/10.1609/aaai.v38i21.30361.
Texte intégralRésumé :
Large language models like ChatGPT can generate human-like code, posing challenges for programming education as students may be tempted to misuse them on assignments. However, there are currently no robust detectors designed specifically to identify AI-generated code. This is an issue that needs to be addressed to maintain academic integrity while allowing proper utilization of language models. Previous work has explored different approaches to detect AI-generated text, including watermarks, feature analysis, and fine-tuning language models. In this paper, we address the challenge of determining whether a student's code assignment was generated by a language model. First, our proposed method identifies AI-generated code by leveraging targeted masking perturbation paired with comperhesive scoring. Rather than applying a random mask, areas of the code with higher perplexity are more intensely masked. Second, we utilize a fine-tuned CodeBERT to fill in the masked portions, producing subtle modified samples. Then, we integrate the overall perplexity, variation of code line perplexity, and burstiness into a unified score. In this scoring scheme, a higher rank for the original code suggests it's more likely to be AI-generated. This approach stems from the observation that AI-generated codes typically have lower perplexity. Therefore, perturbations often exert minimal influence on them. Conversely, sections of human-composed codes that the model struggles to understand can see their perplexity reduced by such perturbations. Our method outperforms current open-source and commercial text detectors. Specifically, it improves detection of code submissions generated by OpenAI's text-davinci-003, raising average AUC from 0.56 (GPTZero baseline) to 0.87 for our detector.
17
Alshammari, Hamed, Ahmed El-Sayed et Khaled Elleithy. « AI-Generated Text Detector for Arabic Language Using Encoder-Based Transformer Architecture ». Big Data and Cognitive Computing 8, no 3 (18 mars 2024) : 32. http://dx.doi.org/10.3390/bdcc8030032.
Texte intégralRésumé :
The effectiveness of existing AI detectors is notably hampered when processing Arabic texts. This study introduces a novel AI text classifier designed specifically for Arabic, tackling the distinct challenges inherent in processing this language. A particular focus is placed on accurately recognizing human-written texts (HWTs), an area where existing AI detectors have demonstrated significant limitations. To achieve this goal, this paper utilized and fine-tuned two Transformer-based models, AraELECTRA and XLM-R, by training them on two distinct datasets: a large dataset comprising 43,958 examples and a custom dataset with 3078 examples that contain HWT and AI-generated texts (AIGTs) from various sources, including ChatGPT 3.5, ChatGPT-4, and BARD. The proposed architecture is adaptable to any language, but this work evaluates these models’ efficiency in recognizing HWTs versus AIGTs in Arabic as an example of Semitic languages. The performance of the proposed models has been compared against the two prominent existing AI detectors, GPTZero and OpenAI Text Classifier, particularly on the AIRABIC benchmark dataset. The results reveal that the proposed classifiers outperform both GPTZero and OpenAI Text Classifier with 81% accuracy compared to 63% and 50% for GPTZero and OpenAI Text Classifier, respectively. Furthermore, integrating a Dediacritization Layer prior to the classification model demonstrated a significant enhancement in the detection accuracy of both HWTs and AIGTs. This Dediacritization step markedly improved the classification accuracy, elevating it from 81% to as high as 99% and, in some instances, even achieving 100%.
18
Kim, Hong Jin, Jae Hyuk Yang, Dong-Gune Chang, Lawrence G. Lenke, Javier Pizones, René Castelein, Kota Watanabe et al. « Assessing the Reproducibility of the Structured Abstracts Generated by ChatGPT and Bard Compared to Human-Written Abstracts in the Field of Spine Surgery : Comparative Analysis ». Journal of Medical Internet Research 26 (26 juin 2024) : e52001. http://dx.doi.org/10.2196/52001.
Texte intégralRésumé :
Background Due to recent advances in artificial intelligence (AI), language model applications can generate logical text output that is difficult to distinguish from human writing. ChatGPT (OpenAI) and Bard (subsequently rebranded as “Gemini”; Google AI) were developed using distinct approaches, but little has been studied about the difference in their capability to generate the abstract. The use of AI to write scientific abstracts in the field of spine surgery is the center of much debate and controversy. Objective The objective of this study is to assess the reproducibility of the structured abstracts generated by ChatGPT and Bard compared to human-written abstracts in the field of spine surgery. Methods In total, 60 abstracts dealing with spine sections were randomly selected from 7 reputable journals and used as ChatGPT and Bard input statements to generate abstracts based on supplied paper titles. A total of 174 abstracts, divided into human-written abstracts, ChatGPT-generated abstracts, and Bard-generated abstracts, were evaluated for compliance with the structured format of journal guidelines and consistency of content. The likelihood of plagiarism and AI output was assessed using the iThenticate and ZeroGPT programs, respectively. A total of 8 reviewers in the spinal field evaluated 30 randomly extracted abstracts to determine whether they were produced by AI or human authors. Results The proportion of abstracts that met journal formatting guidelines was greater among ChatGPT abstracts (34/60, 56.6%) compared with those generated by Bard (6/54, 11.1%; P<.001). However, a higher proportion of Bard abstracts (49/54, 90.7%) had word counts that met journal guidelines compared with ChatGPT abstracts (30/60, 50%; P<.001). The similarity index was significantly lower among ChatGPT-generated abstracts (20.7%) compared with Bard-generated abstracts (32.1%; P<.001). The AI-detection program predicted that 21.7% (13/60) of the human group, 63.3% (38/60) of the ChatGPT group, and 87% (47/54) of the Bard group were possibly generated by AI, with an area under the curve value of 0.863 (P<.001). The mean detection rate by human reviewers was 53.8% (SD 11.2%), achieving a sensitivity of 56.3% and a specificity of 48.4%. A total of 56.3% (63/112) of the actual human-written abstracts and 55.9% (62/128) of AI-generated abstracts were recognized as human-written and AI-generated by human reviewers, respectively. Conclusions Both ChatGPT and Bard can be used to help write abstracts, but most AI-generated abstracts are currently considered unethical due to high plagiarism and AI-detection rates. ChatGPT-generated abstracts appear to be superior to Bard-generated abstracts in meeting journal formatting guidelines. Because humans are unable to accurately distinguish abstracts written by humans from those produced by AI programs, it is crucial to exercise special caution and examine the ethical boundaries of using AI programs, including ChatGPT and Bard.
19
Wani, Mudasir Ahmad, Mohammed ElAffendi et Kashish Ara Shakil. « AI-Generated Spam Review Detection Framework with Deep Learning Algorithms and Natural Language Processing ». Computers 13, no 10 (12 octobre 2024) : 264. http://dx.doi.org/10.3390/computers13100264.
Texte intégralRésumé :
Spam reviews pose a significant challenge to the integrity of online platforms, misleading consumers and undermining the credibility of genuine feedback. This paper introduces an innovative AI-generated spam review detection framework that leverages Deep Learning algorithms and Natural Language Processing (NLP) techniques to identify and mitigate spam reviews effectively. Our framework utilizes multiple Deep Learning models, including Convolutional Neural Networks (CNNs), Long Short-Term Memory (LSTM) networks, Gated Recurrent Unit (GRU), and Bidirectional LSTM (BiLSTM), to capture intricate patterns in textual data. The system processes and analyzes large volumes of review content to detect deceptive patterns by utilizing advanced NLP and text embedding techniques such as One-Hot Encoding, Word2Vec, and Term Frequency-Inverse Document Frequency (TF-IDF). By combining three embedding techniques with four Deep Learning algorithms, a total of twelve exhaustive experiments were conducted to detect AI-generated spam reviews. The experimental results demonstrate that our approach outperforms the traditional machine learning models, offering a robust solution for ensuring the authenticity of online reviews. Among the models evaluated, those employing Word2Vec embeddings, particularly the BiLSTM_Word2Vec model, exhibited the strongest performance. The BiLSTM model with Word2Vec achieved the highest performance, with an exceptional accuracy of 98.46%, a precision of 0.98, a recall of 0.97, and an F1-score of 0.98, reflecting a near-perfect balance between precision and recall. Its high F2-score (0.9810) and F0.5-score (0.9857) further highlight its effectiveness in accurately detecting AI-generated spam while minimizing false positives, making it the most reliable option for this task. Similarly, the Word2Vec-based LSTM model also performed exceptionally well, with an accuracy of 97.58%, a precision of 0.97, a recall of 0.96, and an F1-score of 0.97. The CNN model with Word2Vec similarly delivered strong results, achieving an accuracy of 97.61%, a precision of 0.97, a recall of 0.96, and an F1-score of 0.97. This study is unique in its focus on detecting spam reviews specifically generated by AI-based tools rather than solely detecting spam reviews or AI-generated text. This research contributes to the field of spam detection by offering a scalable, efficient, and accurate framework that can be integrated into various online platforms, enhancing user trust and the decision-making processes.
20
Al Karkouri, Adnane, Fadoua Ghanimi et Salmane Bourekkadi. « Automatic Detection of Generated Texts and Energy : Exploring the Relationship ». E3S Web of Conferences 412 (2023) : 01101. http://dx.doi.org/10.1051/e3sconf/202341201101.
Texte intégralRésumé :
The proliferation of artificial intelligence (AI) and natural language processing (NLP) technologies has enabled the generation of realistic and coherent texts, but it also raises concerns regarding the potential misuse of these technologies for generating misleading or malicious content. Automatic detection of generated texts is crucial in addressing this issue. This article provides a comprehensive examination of the relationship between the detection of generated texts and energy consumption, delving into the techniques, challenges, and opportunities for developing energyefficient algorithms for text detection.
21
Al Karkouri, Adnane, Fadoua Ghanimi et Salmane Bourekkadi. « Unveiling the Environmental Implications of Automatic Text Generation and the Role of Detection Systems ». E3S Web of Conferences 412 (2023) : 01102. http://dx.doi.org/10.1051/e3sconf/202341201102.
Texte intégralRésumé :
The emergence of artificial intelligence (AI) and natural language processing (NLP) technologies has led to the proliferation of automated systems capable of generating text. While these advancements have enhanced various fields, such as language translation and content generation, they have also given rise to concerns regarding the potential misuse of generated texts, particularly in the context of environmental preservation. This scientific article investigates the intricate relationship between automatic detection of generated texts and the environment. We examine the impact of generated texts on environmental awareness, misinformation propagation, and the role of automated detection systems in mitigating the risks associated with generated content. Our findings highlight the crucial need for robust detection mechanisms to preserve the integrity of environmental discourse and ensure sustainable decision-making.
22
Lyu, Siwei. « Wrestling with the deepfakes : Detection and beyond ». Open Access Government 43, no 1 (8 juillet 2024) : 272–73. http://dx.doi.org/10.56367/oag-043-11545.
Texte intégralRésumé :
Wrestling with the deepfakes: Detection and beyond Siwei Lyu, SUNY Empire Innovation Professor from the University at Buffalo, State University of New York, delves into detection and beyond in the realm of DeepFakes, starting with a look at what they are. Since 2017, the term “DeepFake” has become widely known, often appearing in news and media. It combines “deep learning” (a type of artificial intelligence [AI] model) and “fake” to describe synthetic media – like text, audio, images, and videos – created with advanced AI technologies. The concept gained notoriety when a Reddit user began sharing AI-generated pornographic videos that superimposed celebrity faces onto other figures.
23
Fu, Yu, Deyi Xiong et Yue Dong. « Watermarking Conditional Text Generation for AI Detection : Unveiling Challenges and a Semantic-Aware Watermark Remedy ». Proceedings of the AAAI Conference on Artificial Intelligence 38, no 16 (24 mars 2024) : 18003–11. http://dx.doi.org/10.1609/aaai.v38i16.29756.
Texte intégralRésumé :
To mitigate potential risks associated with language models (LMs), recent AI detection research proposes incorporating watermarks into machine-generated text through random vocabulary restrictions and utilizing this information for detection. In this paper, we show that watermarking algorithms designed for LMs cannot be seamlessly applied to conditional text generation (CTG) tasks without a notable decline in downstream task performance. To address this issue, we introduce a simple yet effective semantic-aware watermarking algorithm that considers the characteristics of conditional text generation with the input context. Compared to the baseline watermarks, our proposed watermark yields significant improvements in both automatic and human evaluations across various text generation models, including BART and Flan-T5, for CTG tasks such as summarization and data-to-text generation. Meanwhile, it maintains detection ability with higher z-scores but lower AUC scores, suggesting the presence of a detection paradox that poses additional challenges for watermarking CTG.
24
Gupta, Varun, et Chetna Gupta. « Navigating the Landscape of AI-Generated Text Detection : Issues and Solutions for Upholding Academic Integrity ». Computer 57, no 11 (novembre 2024) : 118–23. http://dx.doi.org/10.1109/mc.2024.3445068.
Texte intégral
25
Fariello, Serena, Giuseppe Fenza, Flavia Forte, Mariacristina Gallo et Martina Marotta. « Distinguishing Human From Machine : A Review of Advances and Challenges in AI-Generated Text Detection ». International Journal of Interactive Multimedia and Artificial Intelligence In press, In press (2024) : 1. https://doi.org/10.9781/ijimai.2024.12.002.
Texte intégral
26
Gosling, Samuel David, Kate Ybarra et Sara K. Angulo. « A widely used Generative-AI detector yields zero false positives ». Aloma : Revista de Psicologia, Ciències de l'Educació i de l'Esport 42, no 2 (5 décembre 2024) : 31–43. https://doi.org/10.51698/aloma.2024.42.2.31-43.
Texte intégralRésumé :
The widespread availability of generative-AI using Large Language Models (LLMs) has provided the means for students and others to easily cheat on written assignments – that is, students can use AI to generate text and then submit that work as their own. A variety of technical solutions have been developed to detect such cheating. However, concerns have been raised about the dangers of falsely identifying real students’ responses as having been generated by AI. Here we evaluate a Generative AI detector that comes as an option with Turnitin, a widely used plagiarism-detection platform already in use at many universities. We compare 160 responses written by students in a class assignment with 160 responses generated by ChatGPT instructed to complete the same assignment. The ChatGPT responses were generated by 16 different prompts crafted to mimic those that plausibly might be given by individuals seeking to cheat on an assignment. The AI scores for the AI generated responses were significantly higher than the AI scores for the human-generated responses, which were all zero. Clearly, an arms race is set to develop between technology that facilitates cheating and technology that detects it. However, the present findings demonstrate that it is at least possible to deploy technical solutions in this context. Looking ahead, as various AI-methods become legitimate tools and are more seamlessly integrated into almost every aspect of daily life, it is unlikely that purely technical solutions will suffice. Instead, guidelines surrounding academic integrity will have to adapt to new conceptualizations of academic mastery and creative output.
27
Carrillo, Irene, Cesar Fernandez, M. Asuncion Vicente, Mercedes Guilabert, Alicia Sánchez, Eva Gil, Almudena Arroyo et al. « Detecting and Reducing Gender Bias in Spanish Texts Generated with ChatGPT and Mistral Chatbots : The Lovelace Project ». Proceedings of The Global Conference on Women’s Studies 3, no 1 (10 novembre 2024) : 29–42. http://dx.doi.org/10.33422/womensconf.v3i1.466.
Texte intégralRésumé :
Current Artificial Intelligence (AI) systems can effortlessly and instantaneously generate text, images, songs, and videos. This capability will lead us to a future where a significant portion of available information will be partially or wholly generated by AI. In this context, it is crucial to ensure that AI-generated texts and images do not perpetuate or exacerbate existing gender biases. We examined the behavior of two common AI chatbots, ChatGPT and Mistral, when generating text in Spanish, both in terms of language inclusiveness and perpetuation of traditional male/female roles. Our analysis revealed that both tools demonstrated relatively low gender bias in terms of reinforcing traditional gender roles but exhibited higher gender bias concerning language inclusiveness, at least in the Spanish language. Additionally, although ChatGPT showed lower overall gender bias than Mistral, Mistral provided users with more control to modify its behavior through prompt modifiers. As a final conclusion, while both AIs exhibit some degree of gender bias in their responses, this bias is significantly lower than the gender bias present in their human-authored source materials.
28
Garib, Ali, et Tina A. Coffelt. « DETECTing the anomalies : Exploring implications of qualitative research in identifying AI-generated text for AI-assisted composition instruction ». Computers and Composition 73 (septembre 2024) : 102869. http://dx.doi.org/10.1016/j.compcom.2024.102869.
Texte intégral
29
He, Zhaokai, Ruolong Mao et Yu Liu. « Predictive model on detecting ChatGPT responses against human responses ». Applied and Computational Engineering 44, no 1 (5 mars 2024) : 18–25. http://dx.doi.org/10.54254/2755-2721/44/20230078.
Texte intégralRésumé :
The paper investigates the critical differences between AI-generated text and human responses in terms of linguistic patterns, structure, and content. The research makes use of datasets from HC3, collected in 2023. Our results are that ChatGPT with GPT-3.5 is more likely to use words like conjunctions and combinations of words in conversations compared to humans systematically. Our model has high accuracy in identifying AI-generated answers.
30
Zaman, Asim, Baozhang Ren et Xiang Liu. « Artificial Intelligence-Aided Automated Detection of Railroad Trespassing ». Transportation Research Record : Journal of the Transportation Research Board 2673, no 7 (9 mai 2019) : 25–37. http://dx.doi.org/10.1177/0361198119846468.
Texte intégralRésumé :
Trespassing is the leading cause of rail-related deaths and has been on the rise for the past 10 years. Detection of unsafe trespassing of railroad tracks is critical for understanding and preventing fatalities. Witnessing these events has become possible with the widespread deployment of large volumes of surveillance video data in the railroad industry. This potential source of information requires immense labor to monitor in real time. To address this challenge this paper describes an artificial intelligence (AI) framework for the automatic detection of trespassing events in real time. This framework was implemented on three railroad video live streams, a grade crossing and two right-of-ways, in the United States. The AI algorithm automatically detects trespassing events, differentiates between the type of violator (car, motorcycle, truck, pedestrian, etc.) and sends an alert text message to a designated destination with important information including a video clip of the trespassing event. In this study, the AI has analyzed hours of live footage with no false positives or missed detections yet. This paper and its subsequent studies aim to provide the railroad industry with state-of-the-art AI tools to harness the untapped potential of an existing closed-circuit television infrastructure through the real-time analysis of their data feeds. The data generated from these studies will potentially help researchers understand human factors in railroad safety research and give them a real-time edge on tackling the critical challenges of trespassing in the railroad industry.
31
Thanathamathee, Putthiporn, Siriporn Sawangarreerak, Siripinyo Chantamunee et Dinna Nina Mohd Nizam. « SHAP-Instance Weighted and Anchor Explainable AI : Enhancing XGBoost for Financial Fraud Detection ». Emerging Science Journal 8, no 6 (1 décembre 2024) : 2404–30. https://doi.org/10.28991/esj-2024-08-06-016.
Texte intégralRésumé :
This research aims to enhance financial fraud detection by integrating SHAP-Instance Weighting and Anchor Explainable AI with XGBoost, addressing challenges of class imbalance and model interpretability. The study extends SHAP values beyond feature importance to instance weighting, assigning higher weights to more influential instances. This focuses model learning on critical samples. It combines this with Anchor Explainable AI to generate interpretable if-then rules explaining model decisions. The approach is applied to a dataset of financial statements from the listed companies on the Stock Exchange of Thailand. The method significantly improves fraud detection performance, achieving perfect recall for fraudulent instances and substantial gains in accuracy while maintaining high precision. It effectively differentiates between non-fraudulent, fraudulent, and grey area cases. The generated rules provide transparent insights into model decisions, offering nuanced guidance for risk management and compliance. This research introduces instance weighting based on SHAP values as a novel concept in financial fraud detection. By simultaneously addressing class imbalance and interpretability, the integrated approach outperforms traditional methods and sets a new standard in the field. It provides a robust, explainable solution that reduces false positives and increases trust in fraud detection models. Doi: 10.28991/ESJ-2024-08-06-016 Full Text: PDF
32
Ramalakshmi, S., et G. Asha. « Exploring Generative AI : Models, Applications, and Challenges in Data Synthesis ». Asian Journal of Research in Computer Science 17, no 12 (13 décembre 2024) : 123–36. https://doi.org/10.9734/ajrcos/2024/v17i12533.
Texte intégralRésumé :
Generative AI has emerged as a transformative field within artificial intelligence, enabling the creation of new data that mimics real-world information and expands the boundaries of what machines can autonomously generate. This study discuss the various models of generative AI, focusing on Variational Autoencoders (VAEs), Generative Adversarial Networks (GANs), and Auto-Regressive models, each offering distinct approaches and strengths in data generation. VAEs excel in learning latent representations, making them ideal for applications like anomaly detection and data imputation. GANs, renowned for their high-quality image synthesis, have found extensive use in tasks ranging from text-to-image conversion to super-resolution. Auto-Regressive models, on the other hand, are particularly effective in sequential data generation, such as text generation, music composition, and time series prediction. The paper highlights key applications of these models across diverse domains, including image synthesis, text generation, drug discovery, and simulation tasks in fields like healthcare, finance, and entertainment. Additionally, the study emphasizes the evaluation metrics are also called the comparitive parameters crucial for assessing the performance of generative models, such as perceptual quality metrics, Inception Score (IS), and Fréchet Inception Distance (FID), which provide quantitative insights into the quality and diversity of generated data. This study employs a systematic methodology comprising a comprehensive literature review, strategic search queries, and thematic data synthesis to explore generative AI. Key areas of focus include models (VAE, GAN, auto-regressive, flow-based), applications, evaluation techniques, challenges, and recent advances. The analysis identifies emerging trends, novel methods, and critical gaps in the field. This study also compares the performance of three Gen –AI models along with the comparative parameters like data type, Data Type, Applications, Training Complexity, Output Quality, Interpretability, Limitations, Advantages, Computational Cost and Scalability. Generative AI raises ethical concerns, including biases in training data that perpetuate stereotypes and marginalization. It can be misused for harmful purposes like creating deepfakes or spreading misinformation, impacting trust and privacy. Questions of accountability and ownership arise when AI-generated content infringes on intellectual property or causes harm. Addressing these issues is essential for responsible AI deployment.
33
Jadhav, Anurag. « Twitter Sentiment Analysis on Chatgpt Tweets ». International Journal for Research in Applied Science and Engineering Technology 11, no 11 (30 novembre 2023) : 1310–14. http://dx.doi.org/10.22214/ijraset.2023.56738.
Texte intégralRésumé :
Abstract: This study investigates multilingual sentiment analysis within tweets on ChatGPT, an AI conversational model, employing Support Vector Machines (SVM) and BERT, an advanced language model. It aims to detect and classify emotions, including emoji identification, embedded within diverse messages across multiple languages on Twitter. By leveraging SVM's text classification and BERT's contextual understanding in various languages, the research delves into preprocessing techniques and feature engineering for sentiment analysis, encompassing multilingual and emoji detection. Furthermore, it explores the fusion of traditional SVM methods with BERT's state-of-the-art model for multilingual sentiment analysis, emphasizing emotion and emoji detection in AI-generated content on multilingual social media platforms like Twitter. This research yields insights into the successful detection of multilingual sentiment nuances and emotions, including emoji identification. It offers implications for advancing multilingual sentiment analysis in natural language processing across diverse linguistic contexts.
34
Kirthiga, Mrs N., Miriyala Vamsi Krishna, Venkata Naveen Vadlamudi, Makani Venkata Sai Kiran et Dudekula Hussain. « Sign Language Detection Using Deep Learning ». International Journal for Research in Applied Science and Engineering Technology 12, no 3 (31 mars 2024) : 1328–34. http://dx.doi.org/10.22214/ijraset.2024.58630.
Texte intégralRésumé :
Abstract: Millions of citizens worldwide suffer from deaf and hard of hearing (DHH), a communication impairment that makes speaking difficult and necessitates the use of sign language. This communication gap frequently hampers access to education and opportunities for employment. Although AI-driven technologies have been studied to tackle this problem, no research has specifically looked into the intelligent and automatic translation of American sign gestures to text in low-resource languages (LRL), such as Nigerian languages. We suggest a unique end-to-end system for translating the American Sign Language, or ASL, Our framework uses the "no language left behind" translation model and the Transformer-based model for ASL-to-Text generation. converting the text generated from LRL to English. We assessed the ASL-to-Text system's performance. The people who participated were able to understand the translated text and expressed satisfaction with both the Text-to-LRL and ASL-toText models, according to a qualitative analysis of the framework. Our suggested framework shows how AI-driven technologies can promote inclusivity in sociocultural interactions and education, particularly for individuals with DHH living in low-resource environments. To bridge the gap between the non-sign language and hearing/speech impaired communities, sign language recognition is crucial. Sentence detection is more useful in real-world situations than isolated recognition of words, yet it is also more difficult since isolated signs need to be accurately identified and continuous, high-quality sign data with distinct features needs to be collected. Here, we suggest a wearable system for understanding sign language using a convolutional neural network (CNN) that combines inertial measurement units attached to the body with flexible strain sensors to detect hand postures and movement trajectories. A total of 48 frequently used ASL sign language terms were gathered and utilized to train the CNN model. This resulted in an isolated sign language word recognition accuracy of 95.85%.
35
Baron, Philip. « Are AI detection and plagiarism similarity scores worthwhile in the age of ChatGPT and other Generative AI ? » Scholarship of Teaching and Learning in the South 8, no 2 (2 septembre 2024) : 151–79. http://dx.doi.org/10.36615/sotls.v8i2.411.
Texte intégralRésumé :
Recent advancements in chatbots have provided students and academics with a new mode of how knowledge can be sourced and composed. Within a very short space of time, students and academics have flocked to use ChatGPT and other Generative Artificial Intelligence (GAI) platforms owing to their capable responses. Additionally, apart from the generative chatbots (such as ChatGPT and Gemini), AI writing tools for paraphrasing, summarising, and co-writing have also become capable and increasingly prevalent to such a degree that the public is spoilt for choice. Having conducted tests on popular chatbots and AI writing tools, it became clear that while programs like Turnitin are developing new algorithms to detect plagiarism and AI-generated content, the initial findings from this study suggest that this may be an increasingly difficult task. These tests were published on YouTube, and within a few weeks, the evidence garnered tens of thousands of views as students and educators seemed uncertain about the strengths, weaknesses, and legalities of these AI tools. What is clear is that we have passed the tipping point, and AI assistance is no longer just a grammar fixer. The implications of this are concerning and far-reaching, as plagiarism is already a significant problem in universities. This position paper reports on tests conducted using Turnitin software and AI writing tools such as ChatGPT and QuillBot. These real-world tests support the paper’s position that it is becoming increasingly difficult to determine what constitutes original work in a world of GAI. The aim of this article is to provide evidence that educators who rely on similarity checking and AI detectors in their current form may inadvertently be supporting plagiarism rather than reducing it. A new method of academic plagiarism detection is proposed, utilizing large language models to generate and track ideas, thereby serving as an idea database. The proposed method focuses on the "understanding" of the work rather than on text similarity.
36
Surianarayanan, Chellammal, John Jeyasekaran Lawrence, Pethuru Raj Chelliah, Edmond Prakash et Chaminda Hewage. « Convergence of Artificial Intelligence and Neuroscience towards the Diagnosis of Neurological Disorders—A Scoping Review ». Sensors 23, no 6 (13 mars 2023) : 3062. http://dx.doi.org/10.3390/s23063062.
Texte intégralRésumé :
Artificial intelligence (AI) is a field of computer science that deals with the simulation of human intelligence using machines so that such machines gain problem-solving and decision-making capabilities similar to that of the human brain. Neuroscience is the scientific study of the struczture and cognitive functions of the brain. Neuroscience and AI are mutually interrelated. These two fields help each other in their advancements. The theory of neuroscience has brought many distinct improvisations into the AI field. The biological neural network has led to the realization of complex deep neural network architectures that are used to develop versatile applications, such as text processing, speech recognition, object detection, etc. Additionally, neuroscience helps to validate the existing AI-based models. Reinforcement learning in humans and animals has inspired computer scientists to develop algorithms for reinforcement learning in artificial systems, which enables those systems to learn complex strategies without explicit instruction. Such learning helps in building complex applications, like robot-based surgery, autonomous vehicles, gaming applications, etc. In turn, with its ability to intelligently analyze complex data and extract hidden patterns, AI fits as a perfect choice for analyzing neuroscience data that are very complex. Large-scale AI-based simulations help neuroscientists test their hypotheses. Through an interface with the brain, an AI-based system can extract the brain signals and commands that are generated according to the signals. These commands are fed into devices, such as a robotic arm, which helps in the movement of paralyzed muscles or other human parts. AI has several use cases in analyzing neuroimaging data and reducing the workload of radiologists. The study of neuroscience helps in the early detection and diagnosis of neurological disorders. In the same way, AI can effectively be applied to the prediction and detection of neurological disorders. Thus, in this paper, a scoping review has been carried out on the mutual relationship between AI and neuroscience, emphasizing the convergence between AI and neuroscience in order to detect and predict various neurological disorders.
37
Audi Albtoush, Et al. « ChatGPT : Revolutionizing User Interactions with Advanced Natural Language Processing ». International Journal on Recent and Innovation Trends in Computing and Communication 11, no 9 (5 novembre 2023) : 3354–60. http://dx.doi.org/10.17762/ijritcc.v11i9.9541.
Texte intégralRésumé :
Significant advancements in artificial intelligence (AI) have unexpectedly improved the standard of living globally. One recent development garnering attention is ChatGPT, a natural language processing (NLP) model created by OpenAI. ChatGPT combines OpenAI's GPT-2 language model with supervised and reinforcement learning techniques, leveraging the extensive language patterns in the GPT-3 corpus. This enables natural text-based interactions between users and AI systems, making it suitable for customer service applications and the creation of voice and text-based virtual assistants. ChatGPT offers features such as topic detection, sentiment detection, sentiment analysis, and the ability to generate multiple threads, enhancing users' understanding and enabling realistic interactions. This paper explores the challenges in AI development and proposed strategies to overcome them. It further examines how ChatGPT can enhance sectors such as chat e-commerce, education, entertainment, finance, health, news, and productivity, highlighting current use cases and potential future applications. The article emphasizes ChatGPT's potential to generate personalized content, making it a promising technology for improving user experiences and advancing AI-driven interactions.
38
Ali, Miss Aliya Anam Shoukat. « AI-Natural Language Processing (NLP) ». International Journal for Research in Applied Science and Engineering Technology 9, no VIII (10 août 2021) : 135–40. http://dx.doi.org/10.22214/ijraset.2021.37293.
Texte intégralRésumé :
Natural Language Processing (NLP) could be a branch of Artificial Intelligence (AI) that allows machines to know the human language. Its goal is to form systems that can make sense of text and automatically perform tasks like translation, spell check, or topic classification. Natural language processing (NLP) has recently gained much attention for representing and analysing human language computationally. It's spread its applications in various fields like computational linguistics, email spam detection, information extraction, summarization, medical, and question answering etc. The goal of the Natural Language Processing is to style and build software system which will analyze, understand, and generate languages that humans use naturally, so as that you just could also be ready to address your computer as if you were addressing another person. Because it’s one amongst the oldest area of research in machine learning it’s employed in major fields like artificial intelligence speech recognition and text processing. Natural language processing has brought major breakthrough within the sector of COMPUTATION AND AI.
39
Turgeon, D. Kim, Lena Krammes, Hiba-Tun-Noor Mahmood, Friederike Frondorf, Vanessa Königs, Julia Luther, Christian Schölz et al. « A novel, noninvasive, multimodal screening test for the early detection of precancerous lesions and colorectal cancers using an artificial intelligence–based algorithm. » Journal of Clinical Oncology 42, no 16_suppl (1 juin 2024) : 3627. http://dx.doi.org/10.1200/jco.2024.42.16_suppl.3627.
Texte intégralRésumé :
3627 Background: Colorectal cancer (CRC) ranks as the second leading cause of cancer-related mortality worldwide, and its incidence is increasing in younger populations. Detection of early-stage CRC and its precursor lesions, such as advanced adenomas (AA), is crucial for successful treatment and reduces CRC-related mortality. Although non-invasive methods for early detection are available and increase screening compliance, their performance with respect to detection of advanced adenomas and early-stage cancers is limited. Here, we describe a novel and non-invasive stool-based approach combining diagnostic biomarkers with an algorithm generated by machine learning/artificial intelligence (ML/AI) resulting in significantly improved diagnostic performance for the detection of not only CRC, but especially precancerous lesions like AA. Methods: Data were generated from a combined cohort of stool samples collected at 10 clinical sites in Europe (COLOFUTURE study) and 21 clinical sites in the US (eAArly DETECT study). The evaluable study cohort consists of 690 subjects, including 78 CRC, 146 AA, 147 with non-advanced adenoma (AD) and 319 normal colonoscopy negative control (NC) subjects (49% female, 51% male, average age 61.8 years). Results were compared with colonoscopy and pathology findings to determine sensitivity and specificity for detection of early-stage CRC and AA vs. AD and NC. Nucleic acids were extracted from stabilized stool samples using a silica bead-based extraction method. Expression of mRNA biomarkers was analyzed utilizing Real-Time PCR. Human hemoglobin was quantified using FIT. The Emerge Quantitative Evolution ML/AI platform was leveraged to develop classifiers capable of distinguishing CRC and AA from AD and NC. Results: Applying the combined approach of non-invasive diagnostic testing with an AI/ML generated algorithm, the sensitivity for detection of CRC overall was 92.3%. In addition, this method enabled AA detection with a sensitivity of 82.2%. Specificity turned out to be 90.1 % (AD+NC). Conclusions: This innovative, non-invasive, multimodal screening strategy based on self-collected stool samples which combines mRNA expression patterns and FIT analysis with an AI/ML-generated algorithm represents a substantial improvement in the effectiveness of non-invasive CRC screening. Such improvements in usability and performance leading to reliable detection of early-stage CRC and precursor lesions, such as advanced adenomas, are required for wide-spread availability and adaption of non-invasive screening tests to accepted clinically relevant usage and to prevent the development of CRC effectively.
40
Patil, Prof Shital. « Air Handwriting using AI and ML ». INTERANTIONAL JOURNAL OF SCIENTIFIC RESEARCH IN ENGINEERING AND MANAGEMENT 08, no 05 (14 mai 2024) : 1–5. http://dx.doi.org/10.55041/ijsrem33918.
Texte intégralRésumé :
Air-writing refers to virtually writing linguistic characters through hand gestures in three dimensional space with six degrees of freedom. In this paper a generic video camera dependent convolutional neural network (CNN) based air-writing framework has been proposed. Gestures are performed using a marker of fixed color in front of a generic video camera followed by color based segmentation to identify the marker and track the trajectory of marker tip. A pre-trained CNN is then used to classify the gesture. The recognition accuracy is further improved using transfer learning with the newly acquired data. The performance of the system varies greatly on the illumi nation condition due to color based segmentation. In a less fluctuating illumination condition the system is able to recognize isolated unistroke numerals of multiple languages. The proposed framework achieved 97.7recognition rate in person inde pendent evaluation over English, Bengali and Devanagari numerals, respectively. Object tracking is considered as an important task within the field of Computer Vision. The invention of faster computers, availability of inexpensive and good quality video cameras and demands of automated video analysis has given popularity to object tracking techniques. Generally, video analysis procedure has three major steps: firstly, detecting of the object, secondly tracking its movement from frame to frame and lastly analysing the behaviour of that object. For object tracking, four different issues are taken into account; selection of suitable object representation, feature selection for tracking, object detection and object tracking. In real world, Object tracking algorithms are the primarily part of different applications such as: automatic surveillance, video indexing and vehicle navigation etc. The generated text can also be used for various purposes, such as sending messages, emails, etc. It will be a powerful means of communication for the deaf. It is an effective communication method that reduces mobile and laptop usage by eliminating the need to write. Key Words: Air Writing, Character Recognition, Object Detection, Real-Time Gesture Control System, Computer Vision , Hand tracking.
41
Loh, Peter K. K., Aloysius Z. Y. Lee et Vivek Balachandran. « Towards a Hybrid Security Framework for Phishing Awareness Education and Defense ». Future Internet 16, no 3 (1 mars 2024) : 86. http://dx.doi.org/10.3390/fi16030086.
Texte intégralRésumé :
The rise in generative Artificial Intelligence (AI) has led to the development of more sophisticated phishing email attacks, as well as an increase in research on using AI to aid the detection of these advanced attacks. Successful phishing email attacks severely impact businesses, as employees are usually the vulnerable targets. Defense against such attacks, therefore, requires realizing defense along both technological and human vectors. Security hardening research work along the technological vector is few and focuses mainly on the use of machine learning and natural language processing to distinguish between machine- and human-generated text. Common existing approaches to harden security along the human vector consist of third-party organized training programmes, the content of which needs to be updated over time. There is, to date, no reported approach that provides both phishing attack detection and progressive end-user training. In this paper, we present our contribution, which includes the design and development of an integrated approach that employs AI-assisted and generative AI platforms for phishing attack detection and continuous end-user education in a hybrid security framework. This framework supports scenario-customizable and evolving user education in dealing with increasingly advanced phishing email attacks. The technological design and functional details for both platforms are presented and discussed. Performance tests showed that the phishing attack detection sub-system using the Convolutional Neural Network (CNN) deep learning model architecture achieved the best overall results: above 94% accuracy, above 95% precision, and above 94% recall.
42
Banat, Maysaa. « Investigating the Linguistic Fingerprint of GPT-4o in Arabic-to-English Translation Using Stylometry ». Journal of Translation and Language Studies 5, no 3 (30 septembre 2024) : 65–83. http://dx.doi.org/10.48185/jtls.v5i3.1343.
Texte intégralRésumé :
This study explores the linguistic and stylistic characteristics of machine-generated texts, focusing on the output of GPT-4o. Using various natural language processing (NLP) techniques, including word frequency and stopword count analysis, readability and sentence structure metrics, lexical diversity measures, syntactic frequency analysis, and named entity recognition (NER), the research aims to uncover the stylometric fingerprints present in machine-generated content. The results reveal that GPT-4ogenerated texts exhibit moderate lexical diversity and syntactic complexity, with certain chapters reflecting higher readability and more varied sentence structures, while others lean toward simpler linguistic patterns. The findings also highlight thematic variation across chapters, as observed in the distribution of named entities, which contributes to understanding the model’s handling of different contextual content. The research suggests that while GPT-4o maintains a consistent style in its generated text, there are distinguishable characteristics that may serve as indicators of machine authorship. This provides valuable insights for stylometric analysis, authorship attribution, and the identification of machine-generated texts in various contexts. Future research could extend this work by exploring deeper stylometric features, conducting cross-model comparisons, and developing advanced authorship detection algorithms tailored for AI-generated content. Moreover, the ethical implications of stylometric analysis in the context of AI-generated texts warrant further investigation, particularly as machine-generated content becomes increasingly prevalent across different domains.
43
Wang, Fan, Afeng Wang, Minghao Pan, Shengli Deng, Qianwen Qian, Ruiqi Jia et Ruyi Zheng. « Recognizing Large‐Scale AIGC on Search Engine Websites Based on Knowledge Integration and Feature Pyramid Network ». Proceedings of the Association for Information Science and Technology 61, no 1 (octobre 2024) : 679–84. http://dx.doi.org/10.1002/pra2.1079.
Texte intégralRésumé :
ABSTRACTThe proliferation of Artificial Intelligence Generated Content (AIGC) poses significant challenges to user experience and information accuracy, especially on search engine websites(Guo et al., 2023). The current solution is to identify AIGC by machine learning algorithms or publicly available AI detection tools, whereas, machine learning(Wang & Wang, 2022) algorithms degrade in accuracy as more data is available and tools such as GPTZero perform poorly in the task of AIGC detection on social media. In this paper, we propose an EPCNN model to identify AIGC on search engine websites, which maintains good performance in large‐scale samples. The ERNIE model integrates cross‐domain knowledge and improves language understanding and generalization. We use ERNIE to extract text features, then use a feature pyramid network to capture semantic information at different levels, and finally use an end‐to‐end structure to connect ERNIE and the feature pyramid network to construct the EPCNN. Experimental results show that our proposed algorithm has high accuracy and the ability to handle large‐scale data compared with machine learning algorithms and AI detection tools.
44
Carvalho, Floran, Julien Henriet, Francoise Greffier, Marie-Laure Betbeder et Dana Leon-Henri. « Deep learning for the detection of acquired and non-acquired skills in students' algorithmic assessments ». Journal of Education and e-Learning Research 10, no 2 (3 février 2023) : 111–18. http://dx.doi.org/10.20448/jeelr.v10i2.4449.
Texte intégralRésumé :
This research is part of the Artificial Intelligence Virtual Trainer (AI-VT) project which aims to create a system that can identify the user's skills from a text by means of machine learning. AI-VT is a case-based reasoning learning support system can generate customized exercise lists that are specially adapted to user needs. To attain this outcome, the relevance of the first proposed exercise must be optimized to assist the system in creating personalized user profiles. To solve this problem, this project was designed to include a preliminary testing phase. As a generic tool, AI-VT was designed to be adapted to any field of learning. The most recent application of AI-VT was in the field of computer science specifically in the context of the fundamentals of algorithmic learning. AI-VT can and will also be useful in other disciplines. Developed in Python with the Keras API and the Tensorflow framework, this artificial intelligence-based tool encompasses a supervised learning environment, multi-label text classification techniques and deep neural networks. This paper presents and compares the performance levels of the different models tested on two different data sets in the context of computer programming and algorithms.
45
Shaik Vadla, Mahammad Khalid, Mahima Agumbe Suresh et Vimal K. Viswanathan. « Enhancing Product Design through AI-Driven Sentiment Analysis of Amazon Reviews Using BERT ». Algorithms 17, no 2 (30 janvier 2024) : 59. http://dx.doi.org/10.3390/a17020059.
Texte intégralRésumé :
Understanding customer emotions and preferences is paramount for success in the dynamic product design landscape. This paper presents a study to develop a prediction pipeline to detect the aspect and perform sentiment analysis on review data. The pre-trained Bidirectional Encoder Representation from Transformers (BERT) model and the Text-to-Text Transfer Transformer (T5) are deployed to predict customer emotions. These models were trained on synthetically generated and manually labeled datasets to detect the specific features from review data, then sentiment analysis was performed to classify the data into positive, negative, and neutral reviews concerning their aspects. This research focused on eco-friendly products to analyze the customer emotions in this category. The BERT and T5 models were finely tuned for the aspect detection job and achieved 92% and 91% accuracy, respectively. The best-performing model will be selected, calculating the evaluation metrics precision, recall, F1-score, and computational efficiency. In these calculations, the BERT model outperforms T5 and is chosen as a classifier for the prediction pipeline to predict the aspect. By detecting aspects and sentiments of input data using the pre-trained BERT model, our study demonstrates its capability to comprehend and analyze customer reviews effectively. These findings can empower product designers and research developers with data-driven insights to shape exceptional products that resonate with customer expectations.
46
Rahnemoonfar, Maryam, Jimmy Johnson et John Paden. « AI Radar Sensor : Creating Radar Depth Sounder Images Based on Generative Adversarial Network ». Sensors 19, no 24 (12 décembre 2019) : 5479. http://dx.doi.org/10.3390/s19245479.
Texte intégralRésumé :
Significant resources have been spent in collecting and storing large and heterogeneous radar datasets during expensive Arctic and Antarctic fieldwork. The vast majority of data available is unlabeled, and the labeling process is both time-consuming and expensive. One possible alternative to the labeling process is the use of synthetically generated data with artificial intelligence. Instead of labeling real images, we can generate synthetic data based on arbitrary labels. In this way, training data can be quickly augmented with additional images. In this research, we evaluated the performance of synthetically generated radar images based on modified cycle-consistent adversarial networks. We conducted several experiments to test the quality of the generated radar imagery. We also tested the quality of a state-of-the-art contour detection algorithm on synthetic data and different combinations of real and synthetic data. Our experiments show that synthetic radar images generated by generative adversarial network (GAN) can be used in combination with real images for data augmentation and training of deep neural networks. However, the synthetic images generated by GANs cannot be used solely for training a neural network (training on synthetic and testing on real) as they cannot simulate all of the radar characteristics such as noise or Doppler effects. To the best of our knowledge, this is the first work in creating radar sounder imagery based on generative adversarial network.
47
Allen, Laura, Wenqian Xu, Mariko Nishikitani, Vaishnavi Atul Patil et Dana Bradley. « AGE BIAS IN ARTIFICIAL INTELLIGENCE (AI) : A VISUAL PROPERTIES ANALYSIS OF AI IMAGES OF OLDER VERSUS YOUNGER PEOPLE ». Innovation in Aging 7, Supplement_1 (1 décembre 2023) : 986. http://dx.doi.org/10.1093/geroni/igad104.3168.
Texte intégralRésumé :
Abstract Concerns about age-related biases in Artificial Intelligence (AI)’s design, development, and data generation raise alarms for potential social inequalities. AI art-generators like MidJourney, known for creating distinctive and artistic images, have gained attention recently. However, the critical examination of their potential biases remains limited. Our study aims to investigate age-related bias in MidJourney by comparing how older and younger individuals are visually depicted. Using 19 keywords (as text prompts) tied to domains of potential social exclusion for older adults, we automated the collection of “old person” and “young person” images in MidJourney. The text-to-image prompts yielded 456 AI-generated images that were analyzed with Amazon Rekognition image detection software to transform them to quantitative data. Data were compared statistically using STATA (version 16). The results show that the images of older people are significantly less bright (p< 0.001) and less sharp (p< 0.001) than the images of younger people. Additionally, images of older people have significantly more smile expressions (16%; p< 0.001) and are more likely to be wearing glasses (56%; p< 0.001) compared to images of younger people (6% and 12%, respectively). The findings reveal that MidJourney’s algorithm associates older people with a negative sense of antiquity, irrelevance, and dread. The findings also align with the Stereotype Content Model in which older people are seen as having warm personalities but not as competent or capable as younger people. We interpret these results as age bias that the AI art generator MidJourney has learned and in turn, creates, through image generation.
48
Bagate, Rupali Amit, et Ramadass Suguna. « Sarcasm Detection on Text for Political Domain— An Explainable Approach ». International Journal on Recent and Innovation Trends in Computing and Communication 10, no 2s (31 décembre 2022) : 255–68. http://dx.doi.org/10.17762/ijritcc.v10i2s.5942.
Texte intégralRésumé :
In the era of social media, a large volume of data is generated by applications such as the industrial internet of things, IoT, Facebook, Twitter, and individual usage. Artificial intelligence and big data tools plays an important role in devising mechanisms for handling this vast volume of data as per the required usage of data to form important information from this unstructured data. When the data is publicly available on the internet and social media, it is imperative to treat the data carefully to respect the sentiments of the individuals. In this paper, the authors have attempted to solve three problems for treating the data using AI and data science tools, weighted statistical methods, and explainability of sarcastic comments. The first objective of this research study is sarcasm detection, and the next objective is to apply it to a domain-specific political Reddit dataset. Moreover, the last is to predict sarcastic words using counterfactual explainability. The textare extracted from the self-annotated Reddit corpus dataset containing 533 million comments written in English language, where 1.3 million comments are sarcastic. The sarcasm detection based model uses a weighted average approach and deep learning models to extract information and provide the required output in terms of content classification. Identifying sarcasm from a sentence is very challenging when the sentence has content that flips the polarity of positive sentiment into negative sentiment. This cumbersome task can be achieved with artificial intelligenceand machine learningalgorithms that train the machine and assist in classifying the required content from the sentences to keep the social media posts acceptable to society. There should be a mechanism to determine the extent to which the model's prediction could be relied upon. Therefore, the explination of the prediction is essential. We studied the methods and developed a model for detecting sarcasm and explaining the prediction. Therefore, the sarcasm detection model with explainability assists in identifying the sarcasmfrom the reddit post and its sentiment score to classify given textcorrectly. The F1-score of 75.75% for sarcasm and 80% for the explainability model proves the robustness of the proposed model.
49
Pingua, Bhagyajit, Deepak Murmu, Meenakshi Kandpal, Jyotirmayee Rautaray, Pranati Mishra, Rabindra Kumar Barik et Manob Jyoti Saikia. « Mitigating adversarial manipulation in LLMs : a prompt-based approach to counter Jailbreak attacks (Prompt-G) ». PeerJ Computer Science 10 (22 octobre 2024) : e2374. http://dx.doi.org/10.7717/peerj-cs.2374.
Texte intégralRésumé :
Large language models (LLMs) have become transformative tools in areas like text generation, natural language processing, and conversational AI. However, their widespread use introduces security risks, such as jailbreak attacks, which exploit LLM’s vulnerabilities to manipulate outputs or extract sensitive information. Malicious actors can use LLMs to spread misinformation, manipulate public opinion, and promote harmful ideologies, raising ethical concerns. Balancing safety and accuracy require carefully weighing potential risks against benefits. Prompt Guarding (Prompt-G) addresses these challenges by using vector databases and embedding techniques to assess the credibility of generated text, enabling real-time detection and filtering of malicious content. We collected and analyzed a dataset of Self Reminder attacks to identify and mitigate jailbreak attacks, ensuring that the LLM generates safe and accurate responses. In various attack scenarios, Prompt-G significantly reduced jailbreak success rates and effectively identified prompts that caused confusion or distraction in the LLM. Integrating our model with Llama 2 13B chat reduced the attack success rate (ASR) to 2.08%. The source code is available at: https://doi.org/10.5281/zenodo.13501821.
50
« An Empirical Study of AI-Generated Text Detection Tools ». Advances in Machine Learning & ; Artificial Intelligence 4, no 2 (20 octobre 2023). http://dx.doi.org/10.33140/amlai.04.02.03.
Texte intégralRésumé :
Since ChatGPT has emerged as a major AIGC model, providing high-quality responses across a wide range of applications (including software development and maintenance), it has attracted much interest from many individuals. ChatGPT has great promise, but there are serious problems that might arise from its misuse, especially in the realms of education and public safety. Several AIGC detectors are available, and they have all been tested on genuine text. However, more study is needed to see how effective they are for multi-domain ChatGPT material. This study aims to fill this need by creating a multi-domain dataset for testing the state-of-the-art APIs and tools for detecting artificially generated information used by universities and other research institutions. A large dataset consisting of articles, abstracts, stories, news, and product reviews was created for this study. The second step is to use the newly created dataset to put six tools through their paces. Six different artificial intelligence (AI) text identification systems, including "GPTkit," "GPTZero," "Originality," "Sapling," "Writer," and "Zylalab," have accuracy rates between 55.29 and 97.0%. Although all the tools fared well in the evaluations, originality was particularly effective across the board.